Nomenclature and Distribution of the Species of the Porcupinefish Family

Home , Chilomycterus, Chilomycterus antillarum, Cyclichthys, Cyclichthys orbicularis, Diodon, Lophodiodon

Memoirs of Museum Victoria 63(1): 77–90 (2006)

ISSN 1447-2546 (Print) 1447-2554 (On-line) http://www.museum.vic.gov.au/memoirs/index.asp

Nomenclature and distribution of the species of the porcupineﬁ sh family Diodontidae (Pisces, Teleostei)

JEFFREY M. LEIS

Ichthyology, Australian Museum, 6 College St, Sydney, NSW 2010, Australia (e-mail: jeffl @austmus.gov.au) Abstract Leis, J.M. 2006. Nomenclature and distribution of the species of the porcupinefi sh family Diodontidae (Pisces, Teleostei). Memoirs of Museum Victoria 63(1): 77–90. The tetraodontiform fi sh family Diodontidae is widely distributed in tropical and temperate marine waters. The family has more than 70 nominal species, over 60% of which were described in the 100 years following Linnaeus. As a consequence, many descriptions are less than detailed, and many types are no longer extant, if they existed at all. The high incidence of synonymy, the many ʻoldʼ descriptions and the wide geographical distributions of the species has led to a great deal of confusion. The present study, based on examination of diodontid holdings in 29 major collections, and including the extant types of all but two of the nominal species, attempts to clarify the nomenclature and distribution of the species of the family. Although some species boundaries are not entirely clear, only 18 or 19 of the nominal species are herein regarded as valid (one as a subspecies). Tentative assignments of species to genera maintain current usage. Final assignments to genera must await a cladistic analysis of relationships within the family. Four species are circumtropical, four species (plus a subspecies) are confi ned to the Atlantic and appear to form a species group, four species are widely distributed in the tropical Indo-Pacifi c, two species are confi ned to tropical Australasia, and three are endemic to temperate Australia. One species described from New Zealand either occurs also in Australia, or is a synonym of an Australian species. Synonymies, a key to the recognized species and a table of the identities of nominal species are provided.

Keywords Tetraodontiformes, burrﬁ sh, Allomycterus, Chilomycterus, Cyclichthys, Dicotylichthys, Diodon, Lophodiodon, Tragulichthys

Introduction This paper lists the senior synonym of each morphologically- defi ned species, followed by the junior synonyms. Brief The porcupinefi sh family Diodontidae contains about justifi cations for synonymies are provided, as are descriptions 19 species in seven or eight genera of warm to temperate seas. of the distribution of the species based on material examined There are about 75 nominal species in the family, and several or on identifi able literature records. In addition to the key of the species have very wide distributions. The species are provided here, regional keys to the species, and illustrations of conspicuous, readily captured, of unusual morphology, and them, can be found in Leis (1986, ref. 5686 – western Indian have been the focus of interest by naturalists from ancient Ocean); Leis (2001, ref. 26318 – western central Pacifi c Ocean: times. Several species have pelagic stages that reach large this key also covers all species in the eastern Indian Ocean); sizes and differ in appearance from the demersal adults. Allen and Robertson (1994, ref. 22193 – eastern Pacifi c Ocean); All this has led to a great deal of confusion as to the number Leis (2003, ref. 27121 – western central Atlantic Ocean); Leis of species, their distributions and the correct names for (in press, eastern central Atlantic Ocean). The key in Leis them. The purpose of this paper is to clarify a number of (2001, ref. 26318) includes all Indo-Pacifi c species and genera nomenclatural issues and the distributions of the species. recognized herein, except the two temperate Australasian Due to their wide distributions, the species are obvious targets species Diodon nicthemerus Cuvier and Allomycterus pilatus for molecular genetic studies, and it is reasonable to expect Whitley (for these, see Kuiter, 1993, ref. 23929, or Gomon et that some taxa that are currently, on the basis of morphology al., 1994, ref. 22532 ). alone, considered to represent a single, widespread species will be subdivided once genetic studies are undertaken. Materials and methods Conversely, some Atlantic taxa have less than clear separations and may eventually be considered conspecifi c. Therefore, Abbreviations of fi ns are as follows: D, dorsal; A, anal; P, it is important to lay some groundwork for these expected pectoral; C, caudal. The spines mentioned are the dermal spines future studies. (i.e., modifi ed scales): fi ns of diodontids lack spines. These 78 Jeffrey M Leis dermal spines have subdermal bases (or roots) that have either Pacifi c genera, three of which are confi ned to Australasia, two approximately opposing bases upon which the exposed contain species that have a mixture of fi xed and erectile spines spine pivots when it elevates, or three (occasionally four) – Allomycterus, Dicotylichthys, Lophodiodon and Tragulichthys broadly more or less equidistantly spaced bases that render the – are usually recognized (see Gomon, 1994, ref. 22532 ; Leis, exposed spine immobile. The exposed portion of the spine 2001, ref. 26318) and are in this paper. Chilomycterus sensu lato varies in length and shape, but erectile spines are generally (about ten species with nearly all dermal spines fi xed and round in cross-section, whereas fi xed spines can vary from immovable) has been more problematical. Tyler (1980, ref. round to compressed in cross-section. See Leis (1978, ref. 5529; 4477) recognized three groupings of Chilomycterus: 1) ʻAtlantic 1986, ref. 5686; or 2001, ref. 26318) for more information on Chilomycterusʼ (fi ve species confi ned to the Atlantic, and called spine morphology. Behind the massive beak-like jaws of by him antennatus, antillarum, mauretanicus, schoepfi i and diodontids is a grinding, or trituration, plate formed by the spinosus); 2) what I call herein ʻCircumtropical Chilomycterusʼ fused premaxillae and dentaries. This plate is often armed with (a circumtropical group considered by Tyler to consist of four transverse plates of teeth, called trituration teeth. species called by him affi nis, atinga, reticulatus and tigrinus); Specifi c information on types is included only if it and 3) ʻIndo-Pacifi c Chilomycterusʼ (considered by Tyler to supplements or corrects information in Eschmeyer (2005; 17 consist only of orbicularis). I agree with Tyler (1980, ref. 4477) Oct 2005 version). To keep the literature cited list to a manageable that nominal species in each of these three groups are length, it includes only references not included in Eschmeyerʼs morphologically more similar to each other than they are to (2005) on-line database (http://www.calacademy.org/research/ species in the other groups. The type species of Chilomycterus ichthyology/catalog/fi shcatmain.asp). The Eschmeyer reference is Diodon reticulatus Linnaeus (1758, ref. 2787); thus, if these number is included with the text citation. Information on Diodon groupings prove to be valid at the generic level, the circumtropical is included in Leis (1978, ref. 5529) and in Leis and Bauchot group becomes Chilomycterus, and I use it in that sense herein. (1984, ref. 12539). For Diodon, only information on species Cyclichthys typically is used for several Indo-Pacifi c species described since 1978 and information updating species (including two not mentioned by Tyler [1980, ref. 4477]), and distributions is included here. the type of Cyclichthys is the Indo-Pacifi c orbicularis; thus, the I examined specimens of diodontids from the following Indo-Pacifi c grouping can be considered Cyclichthys for the institutions (codes after Leviton et al., 1985, ref. 9683): AMS, purposes of this paper. The Atlantic group of species, regarded ANSP (loans based on holdings list), BMNH, BPBM, CAS, by Tyler to be the most phylogenetically basal, is nearly always CSIRO, FAKU, FMNH (loans based on holdings list), FRSKU, included in Chilomycterus. If these Atlantic species were LACM, MCZ, MNHN, NMNZ, NMV, NSMT, NTM, QM, removed from Chilomycterus, the generic name available for RMNH, ROM, RUSI, SAMA, SIO, SMF, UA, USNM, WAM, them is Lyosphaera, based on the unique pelagic stage found in ZMA, ZMB, ZMUC. Distributions are based primarily on at least some members of this group. Unfortunately, the identity museum specimens examined, but are supplemented with of the type species is not clear (on the basis of distribution, reliable literature accounts. schoepfi i seems most likely). Lyosphaera has not been used as a generic name for these fi ve Atlantic species, and until a full Results cladistic analysis is performed on the group, its use is not recommended. For the purposes of this paper, I use Tylerʼs term, Family Diodontidae “Atlantic Chilomycterus”, to identify this grouping. Diagnosis. Small to medium-sized fi shes to 1 m in length, commonly 20–50 cm. Body wide and capable of great infl ation, Chilomycterus (ex Bibron) Brisout de Barneville, 1846 (sensu covered with massive spines that may be quite long; spines with stricto) large bases, or roots, under the skin; long spines usually erectile Chilomycterus (ex Bibron) Brisout de Barneville, 1846 (type and two-rooted, short spines usually fi xed in erect position by species, Diodon reticulatus Linnaeus) their three-rooted bases. Head broad and blunt; gill opening a Cyanichthys Kaup, 1855 (type species is D. coeruleus [non-D. relatively small, vertical slit immediately before pectoral-fi n caeruleus Quoy and Gaimard] Kaup = D. reticulatus Linnaeus, 1758) base; nasal organ usually in small tubes located in front of Diagnosis. All spines fi xed, with long subdermal bases but large eyes; mouth large, wide and terminal; teeth fused to form short or absent external spines (relatively smaller in larger a strong, beak-like crushing structure without a median suture individuals); some spines on top of head with 4 bases; 10 C dividing the upper and lower jaws into left and right halves. rays; 21–23 vertebrae; heavy jaw teeth, but trituration teeth Dorsal and anal fi ns without spines, set far back on body, and, few; no tentacles; nostril in adult an open, cup-shaped organ like caudal fi n, generally rounded; most fi n rays branched; with reticulations; 1 or more spines wholly on dorsal surface of bases of fi ns often thick and fl eshy; no pelvic fi ns. Lateral line caudal peduncle; fi ns spotted; no large blotches on dorsal inconspicuous. No normal scales. surface of head or trunk. Some additional osteological Genera. There is no generally agreed-upon allocation of species characters are given by Tyler (1980, ref. 4477). to the nominal genera, nor is there any cladistic analysis of the The type species of Chilomycterus is Diodon reticulatus family or any subset of it. Most authors recognize Diodon (fi ve Linnaeus (1758, ref. 2787). Cyanichthys coeruleus Kaup (1855, species, revised by Leis, 1978, ref. 5529) for species in which ref. 2571) was based on an unregistered BMNH specimen of nearly all the dermal spines are erectile. Four monotypic Indo- 43 mm SL (see also Günther, 1870, ref. 1995 ). Although Kaup Species of porcupinefish 79

Table 1. Nominal diodontid species and their current identity. Allomycterus jaculiferus McCulloch is included because of confusion about its identity. Bibliographic details can be found in Eschmeyer (2005). (Le (Linnaeus (Linnaeus (Linnaeus Kaup 1855 Kaup 1855 Jordan & Rutter (Linnaeus 1758) (Linnaeus 1758) (Linnaeus 1758) (Walbaum 1792) (Walbaum 1792) (Walbaum (Walbaum 1792) (Walbaum (Walbaum 1792) (Walbaum (Walbaum 1792) (Walbaum 1792) (Walbaum (Leis and Randall (Bloch,1785) (Bloch,1785) i i i i i i fi fi fi fi fi fi Linnaeus 1758 Whitley 1931 Whitley 1931? (Bianconi 1854) Linnaeus 1758 Linnaeus 1758 Linnaeus 1758 Linnaeus 1758 Linnaeus 1758 Cuvier 1818 Brisout de Barneville 1846 Linnaeus 1758 Linnaeus 1758 Linnaeus 1758 Linnaeus 1758? Linnaeus 1758 Linnaeus 1758 Chilomycterus spinosus mauretanicus Danois 1954) Diodon eydouxii Chilomycterus schoep Diodon holocanthus Dicotylichthys punctulatus Diodon nicthemerus Chilomycterus schoep Lophodiodon calori Diodon holocanthus Diodon hystrix Chilomycterus spinosus 1758) Cyclichthys orbicularis Chilomycterus antillarum 1897? Diodon holocanthus Chilomycterus reticulatus Chilomycterus spinosus 1758)? Diodon hystrix Diodon hystrix Chilomycterus schoep Allomycterus pilatus Cyclichthys orbicularis Chilomycterus schoep Allomycterus pilatus Diodon holocanthus Diodon hystrix Diodon hystrix Dicotylichthys punctulatus Chilomycterus reticulatus Diodon holocanthus Chilomycterus reticulatus Chilomycterus schoep Chilomycterus schoep Diodon holocanthus Cyclichthys spilostylus 1982) Diodon hystrix Chilomycterus spinosus 1758) Kotthaus 1979 Cuvier 1818 Cuvier 1818 von Bonde 1923 Cuvier 1818 Smith 1957 Waite 1900 Waite Cuvier 1818 Le Danois 1954 Cuvier 1818 Ayres 1842 Ayres Cuvier 1818 Kaup 1855 Kaup 1855 Bloch 1785 deKay (ex Mitchill) 1842 Linnaeus 1758 Cuvier 1818 Leis & Randall 1982 Cuvier 1818 Jenkins 1904 Atkinson in Bryant 1888 Poey 1875 Lacèpede 1798 Phillipps 1932 Walbaum 1792 Walbaum Cuvier 1818 Linnaeus 1758 Walbaum 1792 Walbaum Cuvier 1818 i Larrañaga 1923 Mitchill 1815 fi Whitley 1931 Ogilby 1910 Curtiss 1938 Lacepède 1798 Nominal speciesmauretanicus meulenii Current identity melanopsis multimaculatus myersi nicthemerus nigrolineatus nigropunctatus novemmaculatus nudifrons orbe orbicularis orbitosus paraholocanthus parcomaculatus torosus totara triedricum pilatus pentodon pilosus plumierii punctatus punctulatus quadradicatus quadrimaculatus verrucosus whiteleyi schoep sexmaculatus spilostylus reticulatus rivulatus spinosissimus spinosus tigrinus (Linnaeus Kaup 1855 (Linnaeus 1758) Jordan & Rutter Jordan & Rutter (Cuvier 1816) (Cuvier 1816) (Cuvier 1816) (Linnaeus 1758) (Linnaeus 1758) (Linnaeus 1758) (Linnaeus 1758) (Cuvier 1818) (Cuvier 1818) (Walbaum 1792) (Walbaum (de Beaufort 1939) (Walbaum 1792) (Walbaum (Walbaum 1792) (Walbaum (Walbaum 1792) (Walbaum (Bloch,1785) i i i i fi fi fi fi Whitley 1931 Linnaeus 1758 Linnaeus 1758 Linnaeus 1758? Cuvier 1818? (Bianconi 1854) Brisout de Barneville 1846 Shaw (ex Lacepède) 1804 Shaw (ex Lacepède) 1804 Shaw (ex Lacepède) 1804 Brisout de Barneville 1846 Linnaeus 1758 Linnaeus 1758 Linnaeus 1758 Linnaeus 1758 able able fi fi Cyclichthys orbicularis Diodon liturosus Diodon nicthemerus Diodon hystrix Chilomycterus antennatus Chilomycterus reticulatus Dicotylichthys punctulatus Chilomycterus reticulatus Diodon eydouxii Chilomycterus schoep Diodon eydouxii Diodon holocanthus Lophodiodon calori Chilomycterus spinosus Chilomycterus antennatus 1758)? Diodon hystrix Chilomycterus reticulatus Chilomycterus schoep Diodon hystrix 1897 Chilomycterus reticulatus Chilomycterus antennatus Chilomycterus antillarum Chilomycterus schoep Diodon liturosus Tragulichthys jaculiferus Tragulichthys Cyclichthys hardenbergi Diodon holocanthus Diodon hystrix Allomycterus pilatus Diodon holocanthus 1897 jaculiferus Tragulichthys Diodon liturosus Chilomycterus antillarum Chilomycterus schoep Chilomycterus reticulatus Mitchill 1815 Klausewitz 1958 Eigenmann 1891 Bleeker (ex Verreaux) 1865 Verreaux) Bleeker (ex Linnaeus 1758 Bloch & Schneider 1801 de Beaufort 1939 Ogilby 1910 Cuvier 1816 deKay 1842 Bloch and Schneider 1801 (non-Cuvier) McCulloch 1921 Cuvier 1818 Jordan & Rutter 1897 Duméril 1855 Hermann 1804 not identi Whitley 1933 Günther 1910 Lema Lucena Saenger & Quoy & Gaimard 1824 Linnaeus 1758 Kaup 1855 Gronow in Gray 1854 Kaup 1855 Shaw (ex Lacepède) 1804 Brisout de Barneville 1846 Metzelaar 1919 Breder 1927 Günther 1910 Steindachner 1866 Bloch & Schneider 1801 Linnaeus 1758 not identi Castelnau 1872 Linnaeus 1758 Bloch 1785 Günther 1870 Bianconi 1854 nis fi caeruleus blochii briareos californiensis bertolettii Oliveria 1979 bleekeri brachiatus diversispinis echinatus eydouxii fuliginosus bocagei calori cornutus digitalis armillatus atinga atringa galapagosensis areolatus antennatus Nominal speciesaf antillarum Current identity jaculiferus maculato-striatus maculatus maculatus maculifer grandoculis hardenbergi holocanthus hystrix jaculiferus geometricus lissogenys liturosus lineatus 80 Jeffrey M Leis identifi ed this specimen as D. coeruleus Quoy and Gaimard 296) was the fi rst author to express a clear opinion, and (an alternate spelling of caeruleus), it is clearly a juvenile, considered D. atringa Linnaeus to be synonymous with D. orbe pelagic specimen of Chilomycterus reticulatus (Linnaeus). Lacepède (1798, ref. 2708). The latter was based on a specimen Kaup (1855, ref. 2571) asserted the specimen was from New from Brazil – no longer extant – that is clearly identifi able as the Guinea, but according to Günther (1870, ref. 1995) the locality Atlantic Diodon spinosus Linnaeus (see below). Le Danois of the specimen “was never known at the British Museum”. A (1959, ref. 12003) considered atringa Linnaeus to be single species (see below). approximately equivalent to Tylerʼs “Atlantic Chilomycterus” group (which includes D. spinosus Linnaeus) with several Chilomycterus reticulatus (Linnaeus, 1758) subspecies roughly equivalent to Tylerʼs species. In contrast, D. reticulatus Linnaeus (1758, ref 2787) is readily Diodon reticulatus Linnaeus, 1758: 334 (India) Diodon echinatus? Linnaeus, 1758: 335 (India) identifi able. Linnaeus based his description on Artediʼs Diodon tigrinus Cuvier, 1818: 127, pl. 6 (Moluccas)* “Ostracion subrotudus…”. Atredi (1738) cited a Willughby Chilomycterus affi nis Günther, 1870: 314 (unknown locality)* (1686) plate of “Orbis muricatus and reticulatus” that is clearly Chilomycterus californiensis Eigenmann, 1891: 1133 (San Pedro, identifi able as reticulatus by its colour, general morphology, California)* spine distribution and spine shape. The name reticulatus has been Euchilomycterus quadradicatus Waite, 1900: 208 (Lord Howe I.)* in regular use as a senior synonym (in addition to the nine Chilomycterus lissogenys Günther, 1910: 476, pl. 179 (Hawaii) 1870–1926 references listed by Fowler, 1936, ref. 6546 and the Chilomycterus galapagosensis Klausewitz, 1958: 82, fi g. 7 >30 post-1985 references listed by Eschmeyer, [2005]: Lowe, (Galapagos Is.)* 1844, ref. 2833; Brisout de Barenville, 1846, ref. 296; Günther, *extant type 1870, ref. 1995; Poey, 1876, ref. 3510; Goode, 1876, ref. 1832, Based on examination of 55 museum specimens, including 1877, ref. 13360; Jordan and Gilbert, 1883, ref. 2476; Eigenmann, extant types, from throughout its range (circumglobal in warm 1885; Poll, 1959, ref. 12014; Tyler, 1980, ref. 4477; Leis, 1981, waters), I can fi nd no morphological differences among the 1984; Leis and Bauchot, 1984, ref. 12539). Some authors, different nominal species or among geographic locations. There apparently following Jordan and Evermann (1898, ref. 2444, is variation in colour pattern similar to that found in the have considered reticulatus to be a junior synonym of atringa, circumtropical Diodon holocanthus (see Leis, 1978, ref. 5529) but none have attempted to justify this view. It is clear from but, like that species, it is not obviously geographically based. Jordan and Evermannʼs description and key that they incorporated Spotting on the body is variable, although usually present to more than one species in their concept of C. atringa, including at some degree, but smaller spots are present on at least some, and least C. reticulatus and C. antennatus. Jordan and Evermann usually all, of the fi ns. The pelagic juvenile phase has a distinctly (1898) described C. atinga as having dark dorsal blotches and a different colour phase from the benthic adult. The species ʻsupraocular cirrusʼ, both features that are lacking in C. reticulatus remains pelagic to about 200 mm SL, thus providing ample (Linnaeus) and in Tylerʼs ʻcircumtropical Chilomycterusʼ, but opportunity for dispersal and maintenance of genetic continuity. present in species of the ʻAtlantic Chilomycterusʼ group. Therefore, I regard Tylerʼs group of ʻcircumglobal Chilomycterusʼ There is a great deal of confusion in the literature as to just to consist of a single species. The rationale for calling this what constitutes C. reticulatus and C. atringa. For much of the species Chilomycterus reticulatus (Linnaeus) follows. 19th century, most authors accepted Blochʼs (1785, ref. 21381) Diodon atringa Linnaeus (1758, ref. 2787) is frequently concept of D. atinga (=D. hystrix Linnaeus, 1758, ref. 2787), regarded as a synonym (often the senior synonym) of this and although Brisout de Barneville (1846, ref. 296) pointed species. This name is often misspelled atinga by authors. This is out that this was in error, the use of D. atinga sensu Bloch clearly incorrect: Linnaeus used the spelling atringa in both his persisted for some years. Séret and Opic (1981) stated without tenth and twelfth editions. Nelson et al. (2004, ref. 27807) reasons that reticulatus was a synonym of C. atringa recently explained why atringa is correct; and Eschmeyer (2005) (Linnaeus), but their illustration of C. atringa shows what used atringa. To avoid confusion with D. atinga Bloch (1785, appears to be C. antennatus (Cuvier) (Séret and Opic kindly ref. 4866), a synonym of D. hystrix, I herein use the spelling provided unpublished dorsal and lateral views of the specimen atringa for the Linnaean species, regardless of the spelling used that strengthen this opinion). Similarly, Tortonese (1973, ref. by any subsequent author. Unfortunately, D. atringa Linnaeus is 7192), without comment, listed reticulatus as a junior synonym not identifi able. There is no type, and Linnaeusʼ description of atringa, but, later, Tortonese (in Whitehead et al., 1986, ref. could apply to any of several species of Chilomycterus or 13677) illustrated as C. atringa a specimen of the eastern Cyclichthys (including C. reticulatus, C. antennatus or C. Atlantic C. spinosus mauretanicus (Le Danois), but with spinosus), and the same is true of Artedi (1738), the only source spotted fi ns, a feature I have not observed in the latter species. cited by Linnaeus, and the pre-Linnaean authors cited by Artedi. Smith-Vaniz et al., 1999 (ref. 25013) listed C. atringa Artedi, 1738 mentioned that the fi ns of his “Ostracion bidens (Linnaeus) as occurring in Bermuda, but had seen no sphaericus…”, upon which Linnaeus based his D. atringa, were specimens, stating that their listing was based on Goodeʼs spotted, but large individuals of C. antennatus (Cuvier, 1816, (1876, ref. 1832; 1877, ref. 13360) and Güntherʼs (1870, ref. ref. 993 ) also have spotted fi ns (see below), so this is not 1995) records of C. reticulatus. Unfortunately, this leads to diagnostic, as is often assumed. Linnaeus (1766, ref. 2786) cited ambiguity because, Smith-Vaniz et al. (1999, ref. 25013) could a plate in Seba (1759, ref. 18716) that represents either C. be interpreted as considering reticulatus a synonym of atinga, reticulatus or C. antennatus. Brisout de Barneville (1846, ref. or as considering that the other authors misidentifi ed their Species of porcupinefish 81

specimens. Some other workers (e.g., Fowler, 1936, ref. 6546; USNM 43860 is in fact the holotype, in spite of Eigenmannʼs Lozano Rey, 1952) have included reticulatus of authors in statement in the original 1891 description that “I did not obtain their synonymies of atringa, but not reticulatus Linnaeus it”. The holotype is in the pelagic colour phase of C. (1758, ref 2787), implying that they questioned othersʼ concept reticulatus. of reticulatus rather than that they considered reticulatus Euchilomycterus quadradicatus Waite (1900, ref. 4558) Linnaeus to be a junior synonym of atringa. from Lord Howe I. was based on a dried specimen – apparently In more recent years, a view has developed among some a beach wash-up – subsequently preserved in ethanol and in American workers that Atlantic individuals of this taxon are poor condition. Although not fi gured by Waite, Whitley (1952) C. atringa, whereas the Indo-Pacifi c individuals are either C. illustrated the holotype (with some artistic license) clearly affi nis (Robins et al., 1991, ref. 14237) or C. reticulatus (Nelson showing the caudal-peduncle spine and four-rooted spines on et al., 2004, ref. 27807), but, again, no justifi cation for this or the head that in combination are diagnostic of Chilomycterus means of distinguishing the two nominal species has ever been reticulatus. presented. Tyler (1980, ref. 4477) tentatively recognized four Chilomycterus lissogenys Günther (1910, ref. 14460) was species in this complex that have, based on his material based on an illustration by Garrett of a Hawaiian fi sh. Although examined and text, different distributions: C. atringa (western Garrett omitted some of the spines on the side of the head, he Atlantic); C. reticulatus (eastern Atlantic and Indo-Pacifi c); C. clearly showed the spine on the caudal peduncle that is tigrinus (western Indian Ocean); and C. affi nis (Eastern characteristic of C. reticulatus. The illustration showed relatively Pacifi c). However, Tyler (1980, ref. 4477) said that C. tigrinus few spots on the body, but heavily spotted fi ns, a condition well may be the young of C. reticulatus (I agree). So confusion within the range of colour variation in this species. about the identity and distribution of these species continues. The description and photo of Chilomycterus galapagosensis In summary, D. atringa Linnaeus is unidentifi able, and the Klausewitz (1958, ref. 12080) are clearly that of C. reticulatus. post-Linnaean use of the name by various authors has been The description of the nostrils alone is diagnostic. Klausewitz inconsistent as to what species was being included: at least distinguished his new species from C. atringa, which he four species and two multi-species groups have been identifi ed described as having a supraorbital cirrus and large dorsal as D. atringa by various authors at various times. In spite of blotches (presumably based on the description of Jordan and the use of C. atringa (usually spelled atinga) by several Evermann [1898, ref. 2244], which was based on more than authors, the name should be regarded as a nomen dubium, and one species), by its lack of these two characteristics. He not used. Diodon reticulatus Linnaeus is clearly identifi able, distinguished it from C. californiensis by colour, but the latter and the use of the name has been remarkably consistent: it is in pelagic-phase colour, whereas C. galapagosensis has should be used for this species. typical, spotted demersal colour. Diodon echinatus Linnaeus (1758, ref 2787) is seemingly equivalent to his Chilomycterus reticulatus (see Leis and Distribution. Circumglobal in warm temperate to tropical waters: Randall, 1982). Linnaeusʼ (1758, ref 2787) description and the W Atlantic – 39oN to 24oS Marcgrave plate to which Artedi (1738) referred could apply to E Atlantic – Madeira (and possibly to Portugal) and Cape any Chilomycterus or Cyclichthys species. Linnaeus (1766, ref. Blanco to Angola 2786) referred to a Seba (1759, ref. 18716) fi gure that is clearly W Indian Ocean – South Africa to Tanzania and Reunion, Diodon hystrix. Gronow (1854, ref. 6828), in his account of Seychelles and Mauritius “Holocanthus echinatus”, cited a Seba (1759, ref. 18716) fi gure E Indian Ocean – Broome, Western Australia to Bali and that is either Chilomycterus reticulatus or C. antennatus, and a Timor Willughby (1686) fi gure that clearly represents C. reticulatus. W and central Pacifi c – Japan to Lord Howe I. and northern The holotype of Diodon tigrinus Cuvier (1818, ref. 18059) New Zealand, to Tuamotos to Hawaii (and in the east Pacifi c is a specimen in the pelagic colour phase of C. reticulatus. The barrier) species was recognized as a synonym of C. reticulatus as long E Pacifi c – San Pedro, California to Chile, Galapagos and ago as Brisout de Barneville (1846, ref. 296). Revillagigedos Chilomycterus affi nis Günther (1870, ref.1995) was based Occurrences of this species are patchy, and many are based on a specimen of unknown locality that is dried and thickly on pelagic juveniles: in particular, adults are unknown from varnished. The holotype has minimal spotting on the body, and broad areas of the Indo-Pacifi c. Pelagic juveniles are frequently the spines, particularly on the head, are distorted by the found poleward of the distribution of adults in areas of strong, taxidermy and insertion of large, blue glass eyes. However, poleward currents. there is nothing outside of the range of C. reticulatus variability If future work indicates that C. reticulatus contains more in this specimen. In the absence of any locality information, it than one geographically distinct species, several names are is unclear why most authors regarded this as a Pacifi c species. available for Indo-Pacifi c populations, but no name is clearly Chilomycterus californiensis was described by Eigenmann based on Atlantic material. Most of the extant types are either (1891, ref. 18744) on the basis of a specimen that he initially dried or fi xed in alcohol, so it may be possible to obtain genetic did not obtain from the fi sherman who captured it “on account data from them that could be helpful. Unfortunately, there are of the unreasonable price asked for it”. However, the fi sh was no Linnaean types that might assist in this regard, and subsequently “procured by the National Museum”, and Linnaeusʼ usage of ʻhabitat in Indiaʼ cannot be taken at face Eigenmann (1892) redescribed and fi gured it. Therefore, value in most cases. 82 Jeffrey M Leis

Cyclichthys Kaup, 1855 Chilomycterus parcomaculatus von Bonde (1923, ref. 521) was based on a specimen that was ʻinadvertently destroyedʼ Cyclichthys Kaup, 1855 (type species Diodon orbicularis Bloch) (S.X. Kannemeyer, personal communication, 1/2/80), but the Diagnosis. All but 1 or 2 spines fi xed; all spines with 3 bases, description and fi gure are diagnostic. except in C. spilostylus which has some spines on top of head Distribution. Indo-west Pacifi c with 4 bases; 9 C rays; 19–20 vertebrae; no tentacles in adults; W Indian Ocean – Capetown, South Africa to Red Sea, nostril in adult a short tube with 2 openings; no spines wholly on Oman and Persian Gulf, Maldives, Reunion. dorsal surface of caudal peduncle; no fi ns spotted; no large E Indian Ocean – Shark Bay, Western Australia to Burma blotches on dorsal surface. Some additional osteological characters W Pacifi c – southern Japan and Sea of Japan to Sydney, are given by Tyler (1980, ref. 4477) for C. orbicularis. Australia and east to Philippines and New Caledonia. The type species of this genus is C. orbicularis (Bloch, 1785, ref. 21381). Kaup (1855, ref. 2571) included two species in his Cyclichthys – orbicularis Bloch, and cornutus Kaup – but Cyclichthys hardenbergi (de Beaufort, 1939) designated neither as type species for the genus. Subsequently, Chilomycterus hardenbergi de Beaufort, 1939: 33–34 (New Bleeker (1865, ref. 416) was apparently the fi rst to designate a Guinea)* type species for Cyclichthys and chose orbicularis Bloch *extant type (Eschmeyer, 2005). Fraser-Brunner (1943, ref. 1495) used Based on examination of 21 museum specimens from throughout Cyclichthys as a subgenus of Chilomycterus. He did not the limited range, including the holotype. There are no real consider C. hardenbergi, but included C. orbicularis, the questions as to the identity of this species: de Beaufortʼs (1939, “Atlantic Chilomycterus” species, and C. echinatus non- ref. 17230) description is diagnostic, and the type is extant. This Linnaeus (= C. spilostylus) in his concept of the subgenus species has one of the more limited ranges within the family. Cyclichthys. Tyler (1980, ref. 4477) pointed out that C. orbicularis had osteological differences from the other Distribution. Indo-Pacifi c diodontids he studied, and placed the species in a group on its North-western Australia to the west coast of Cape York, own. However, he was not able to examine specimens of C. and the south coast of New Guinea. Kailola (1975) also spilostylus or C. hardenbergi. Based on external morphology, recorded it from the Trobirand Islands. it appears that C. orbicularis differs from other species that have been included in Cyclichthys by Fraser-Brunner (1943, Cyclichthys spilostylus (Leis and Randall, 1982) ref. 1495), and there is merit in Tylerʼs placement. If this were Chilomycterus spilostylus Leis and Randall, 1982: 363, fi gs 1, 2 done, then a new genus would probably have to be described (Red Sea)* for hardenbergi and spilostylus, as they do not appear to be *extant types monophyletic with the “Atlantic Chilomycterus” species. Pending a cladistic analysis of relationships in the family, I Based on examination of 23 museum specimens, including the recognize three species in Cyclichthys, which has been standard types. This species was mis-identifi ed as Cyclichthys echinatus practice in recent years. (Linnaeus, 1758, ref. 2787) by some authors (see Leis and Randall, 1982, ref. 8453 ), but echinatus is most likely a synonym Cyclichthys orbicularis (Bloch, 1785) of Chilomycterus reticulatus (Linnaeus). Other than this, there are no real questions as to the identity of this wide-ranging Diodon orbicularis Bloch, 1785: 73, pl. 127 (Jamaica?, Cape of species. The pelagic stage of this species has a tentacle emerging Good Hope & Moluccas)*? from each spine. These tentacles are lost at settlement. Diodon caeruleus Quoy and Gaimard, 1824: 201, pl. 65 (fi g. 5) (North of New Guinea, 142°E, at the Equator)* Distribution. Indo-Pacifi c Chilomycterus parcomaculatus von Bonde, 1923: 38, pl. 9 (fi g. 2) W Indian Ocean – Capetown, South Africa to Gulf of Elat, (Natal, South Africa) Red Sea, Muscat to western India and Mauritius (also a * extant type Med iter ra nea n record f rom Israel by Gola n i (1993), presu mably Based on examination of 88 lots from throughout the range, via Suez Canal) including the extant types. There has been no real question as E Indian Ocean – Northwest Cape, Western Australia to to the identity of this wide-spread and common species. Bali Although Blochʼs (1785, ref. 21381) types might all be lost, W Pacifi c – Southern Japan to Hong Kong, Philippines, his plate showing the arrangement of spines, especially those New Caledonia, and northern Great Barrier Reef. near the mouth, is diagnostic of this species. There is one specimen E Pacifi c – Galapagos (single record, including photograph, of unknown origin in ZMB that may be a syntype of this species by Humann [1997], repeated by Grove and Lavenberg [1997], (Paepke, 1999, ref. 24282), but defi nitive evidence is lacking. ref. 24023). The alleged type locality of Jamaica appears to be in error. The holotype of Diodon caeruleus Quoy and Gaimard “Atlantic Chilomycterus” (1824, ref. 3574) was described and fi gured, and the specimen Lyosphaera Evermann and Kendall, 1898: 131 (type species is extant (see Leis and Bauchot, 1984, ref. 12539 ), leaving no Lyospharea globosa Evermann and Kendall, possibly = Diodon doubt that it is conspecifi c with C. orbicularis (Bloch). schoepfi i Walbaum, 1792) Species of porcupinefish 83

Atinga Le Danois, 1954: 2356. (type species Diodon atringa Chilomycterus spinosus spinosus (Linnaeus, 1758) Linnaeus – see Eschmeyer, 2005, and below). Diodon spinosus Linnaeus, 1758: 335 (India) Diagnosis. All spines fi xed; all spines with 3 bases; 9 C rays; Diodon orbe Lacepède, 1798: 124, pl. 3 (Rio de Janeiro) 19–20 vertebrae; tentacles present on lower jaw and usually Cyclichthys cornutus Kaup, 1855: 231? (unknown locality)* over eye; nostril in adult a short tube with 2 openings; no spines Tetrodon torosus Larrañaga, 1923: 390? (Uruguay) wholly on dorsal surface of caudal peduncle; no fi ns spotted *extant type (except in large C. antennatus); large blotches present on dorsal Based on examination of 24 lots (50–200 mm) from throughout surface. Some additional osteological characters are given by the range, and the extant type of C. cornutus Kaup (BMNH Tyler (1980, ref. 4477). At least 2 of the species of this group 1849.1.15.36). share the “Lyosphaera” larval stage (antennatus and the type Linnaeus (1758, ref. 2787) based his description of Diodon species of Lyosphaera), and others may do the same. spinosus on Artedi (1738), who cited an illustration by Although the genus Lyosphaera Evermann and Kendall Willughby (1686). This information is suffi cient to determine (1898, ref. 1281) is available as a generic name for this group, I that the species is one of the “Atlantic Chilomycterus” species, recommend against its use until a cladistic analysis of the but without any lines or small spots on the body. This eliminates “Atlantic Chilomycterus” species and their relationship to other schoepfi i, antennatus, antillarum and mauretanicus, leaving diodontids is undertaken. Lyosphaera has never been used in spinosus as the unlined, unspotted species of this group. this way, and the identity of the type species is unclear (although Lacepède (1798, ref. 2708) provided a fi gure of D. orbe most likely to be schoepfi i based on distribution). The species that clearly shows the arrangement of spines and the diagnostic upon which Atinga Le Danois (1954, ref. 6451) is based is dorsal blotches and lack of small spots or lines on the body. unclear. The type species, D. atringa Linnaeus (1758, ref. 2787), This and the type locality leave no doubt that Diodon orbe is is not identifi able (see above under C. reticulatus). It is clear that conspecifi c with C. spinosus (Linnaeus). Le Danoisʼ (1954, ref. 6451; 1959, ref. 12003; 1962, 21440) The type of Cyclichthys cornutus Kaup (1855, ref. 2571) is concept of atinga included a species of the “Atlantic a small, stuffed specimen of unknown origin with a thick coat of varnish, but the spine arrangement and presence of a Chilomycterus” group, although which species is unclear as her supraocular tentacle show that it is clearly a species of the illustrations of Atinga atinga atinga in the 1954 paper are of C. “Atlantic Chilomycterus” group. The visible colour pattern antillarum (identifi ed as male) and C. spinosus mauretanicus best fi ts C. spinosus (Linnaeus). (identifi ed as female). In view of this confusion about the identity Tetrodon torosus Larrañaga (1923: 390 ref. 22561: not seen of the type species, use of Atinga Le Danois (1954: ref. 6451) is by me) has been regarded as a synonym of Chilomycterus spinosus not recommended. It has been little used since its description. (Linnaeus) since 1925 (Devincenzi, 1925, ref. 20322, see The “Atlantic Chilomycterus” is a group of similar species Eschmeyer, 2005) and I am unaware of any subsequent use of the previously recognized in various ways by Günther (1870, ref. name. If the synonymy of Devincenzi is correct, Uruguay would 1995), Le Danois (1959, ref. 12003) and Tyler (1980, ref. 4477). represent the southernmost record of C. spinosus spinosus. See above regarding the generic status of these species. As noted above, eastern Atlantic specimens of C. spinosus Chilomycterus antennatus is the only member of this group have oblique, irregular lines laterally on the trunk and head that I can separate on morphological grounds; principally, the that are lacking in western Atlantic specimens. Therefore, I development of the fl eshy tentacles over the eye. It also has a have recognized the western Atlantic population as the colour pattern that differs more from the other species of the nominate subspecies and the eastern Atlantic population as C. “Atlantic Chilomycterus” group than they do from each other. spinosus mauretanicus (Le Danois) (see below). The other four taxa differ only in colour, and have largely non- overlapping distributions. In all but the case of the very similar Distribution. Western Atlantic forms, C. spinosus (Linnaeus, 1758, ref. 2787) and C. From northern coast of South America (Surinam and British Guiana) to Rio de Janeiro, Brazil. mauretanicus (Le Danois, 1954, ref. 6451), the distributions do at least seem to come into contact. In contrast, the latter Chilomycterus spinosus mauretanicus (Le Danois, 1954) two taxa occur only on opposite sides of the Atlantic and they have only very minor differences in colouration. Hence, I treat Atinga atinga mauretanicus Le Danois, 1954: 2354 (Mauritania, these two as subspecies: Chilomycterus spinosis spinosus and Gulf of Guinea)* Chilomycterus spinosus mauretanicus. In some cases, colour *extant types patterns do exhibit intermediacy. Chilomycterus schoepfi i Based on examination of 21 lots from throughout the range, adults have a distinctive lined pigment pattern, but the youngest including the syntypes. This nominal species is considered a C. schoepfi i have a colour pattern not unlike that of C. spinosus subspecies of C. spinosus (Linnaeus) because only minor (dark background with lighter, diffuse spotting), and at colour differences separate it from its western Atlantic intermediate sizes, the dark background may have shrunk to a counterpart. Le Danois (1954, ref. 6451) briefl y described this mesh-like pattern with expanded lighter centres similar to that nominal species, apparently inadvertently, in a paper on sexual of C. antillarum. Similarly, in northern South America, a dimorphism in diodontids, then redescribed it in 1959 (ref. colour pattern with elements of both C. spinosus and C. 12003) and provided more information – some of it confl icting antillarum is present. Examination of the genetics of these – in 1962 (ref. 21440 ). See Leis and Bauchot (1984, ref. 12539) “Atlantic Chilomycterus” species would be very interesting. for information on the status of the types. 84 Jeffrey M Leis

Distribution. Eastern Atlantic. of the types of D. antennatus Cuvier [1816, ref. 993]). In spite of From central Angola to Canary Is. and perhaps Portugal. assertions to the contrary, C. antennatus can have spotted fi ns. Fin spotting in C. antennatus begins basally on all fi ns at about Chilomycterus schoepfi i (Walbaum, 1792) 50 mm SL. The caudal fi n becomes mostly or entirely spotted by 100–150 mm SL. Spotting on other fi ns seems variable, but Diodon schoepfi i Walbaum, 1792: 601 (New York) Diodon meulenii Walbaum, 1792: 602 (unknown locality) basal one-third to one-half of the P, D and A fi ns can be spotted Diodon geometricus var. lineatus Bloch and Schneider, 1801: 513 in specimens as small as 127 mm SL, whereas other specimens (New York) as large as 200 mm may have spots only on the extreme base on Diodon maculato-striatus Mitchill, 1815: 470 (New York) these fi ns. Because many ichthyologists have assumed that any Diodon rivulatus Cuvier, 1818: 129, pl. 6 (unknown locality [New Chilomycterus with spotted fi ns is C. reticulatus (or one of its York, USA according to Eschmeyer, 2005])* synonyms), this has led to many misidentifi cations of C. Diodon nigrolineatus Ayres, 1842: 68 (Brookhaven, Long Island, antennatus, and is probably the basis for Jordan and Evermannʼs New York) (1998, ref. 2244) inclusion of what are apparently characteristics Diodon fuliginosus deKay, 1842: 324, pl. 55 (fi g. 181) (New York) of C. antennatus in their description of C. atringa. Diodon verrucosus deKay (ex Mitchill), 1842: 325, pl. 55 (fi g. 1)? Aside from colour differences, C. antennatus has larger (New York) Holocanthus areolatus Gronow in Gray, 1854: 27? (Cape of Good fl eshy tentacles, particularly over the eye, than do the other Hope, South Africa?) “Atlantic Chilomycterus” species. It clearly has a Lyosphaera Chilomycterus pentodon Atkinson in Bryant, 1888: 18 (Beaufort, stage larva (Heck and Weinstein, 1978). North Carolina, USA) Cuvierʼs (1816, ref. 993) description and fi gure were * extant type. diagnostic of the species, and what is probably the type is extant in MNHN (see Leis and Bauchot, 1984, ref. 12539), Based on examination of 62 lots from throughout the range. leaving no doubt about the identity of this distinctive species. Unfortunately, I could locate types of only one of the ten nominal The description and fi gure of Chilomycterus briareos species represented here. Metzelaar (1919 ref. 2982) clearly refers to C antennatus, and Walbaumʼs (1792, ref. 4572) description of D. schoepfi i the type is extant. The fi sh has spots on the fi ns, particularly mentions the diagnostic lined colour pattern of this species, as do on the caudal fi n, which is common in larger individuals of C. the descriptions of Diodon meulenii Walbaum (1792, ref. 4572), antennatus. Diodon geometricus var lineatus Bloch and Schneider (1801, ref. In contrast, Lyosphaera digitalis Breder (1927, ref. 635), is 471), Diodon maculato-striatus Mitchill (1815, ref. 13292), the young ʻLyosphaera stageʼ of this species, virtually lacking Diodon rivulatus Cuvier (1818, ref. 18059), Diodon nigrolineatus spines. Heck and Weinstein (1978) have documented the Ayres (1842, ref. 15926), Diodon fuliginosus deKay (1842, ref. transition of this distinctive ʻLyosphaera stageʼ to the juvenile 1098), Holocanthus areolatus Gronow in Gray (1854, ref. 6828) of C. antennatus. and Chilomycterus pentodon Atkinson in Bryant (1888, ref. 13034), thus confi rming their identifi cation. The type locality of Distribution. Western Atlantic (possibly eastern Atlantic). South Africa for H. areolatus provided by Gronow introduces W Atlantic – Key West Florida to Panama, Colombia and some doubt, but this may well be an error, as there is no diodontid Tobago, Bermuda, and throughout Caribbean and Antilles. species with a lined colour pattern in that area. Diodon E Atlantic – no specimens, but see below. verrucosus deKay (ex Mitchill) (1842, ref. 1098) has a pigment There are persistent reports of C. antennatus from the pattern similar to that of C. antillarum, but, apparently, C. eastern Atlantic, but I have seen no specimens from this area. schoepfi i passes through an early life-history phase with this Where published descriptions or illustrations of “Chilomycterus colour pattern, and the type locality of New York would seem to antennatus” from the eastern Atlantic are diagnostic, they are eliminate the tropical C. antillarum, so I tentatively consider usually of C. spinosus mauretanicus, or in some cases C. verrucosus to be a synonym of C. schoepfi i. reticulatus. However, there is one published illustration of a fi sh from Senegal that does appear to be C. antennatus, although Distribution. Western North Atlantic it is identifi ed in the publication as C. atringa (Linnaeus) (Séret From Halifax, Nova Scotia (waif) to Belize (apparently and Opic, 1981). Unfortunately, the specimen was not retained with a gap between southern Texas and Belize) on the mainland (B. Séret, personal communication). When I requested a and Cuba, Bermuda and Bahamas. specimen for study, Séret, who was not in Senegal at the time, kindly arranged for a colleague to send me one: it was C. Chilomycterus antennatus (Cuvier, 1816) spinosus mauretanicus. Therefore, it is possible that C. Diodon antennatus Cuvier, 1816: 185, pl. 9 (unknown locality)*? antennatus does occur rarely in the eastern Atlantic, most likely Chilomycterus briareos Metzelaar, 1919: 173, fi g. 55 (Lesser as a waif from the west. Specimens are needed to confi rm this. Antilles, St Eustatius)* Lyosphaera digitalis Breder, 1927: 81, fi g. 34 (locality unknown Chilomycterus antillarum Jordan and Rutter, 1897 Western North Atlantic or W Indies)* *extant type. Diodon geometricus Bloch and Schneider, 1801: 513, pl. 96 (coast of Brazil) Based on examination of 37 lots, including extant types (see Chilomycterus antillarum Jordan and Rutter, 1897: 131 (Kingston, Leis and Bauchot, 1984, ref. 12539, for a discussion of the status Jamaica)* Species of porcupinefish 85

* extant type additional osteological characters are given by Tyler (1980, ref. 4477) as Diodon jaculiferus. Based on examination of 41 specimens including extant types. The type species of this monotypic genus is D. jaculiferus Although D. geometricus Bloch and Schneider (1801, ref. 471) Cuvier (1818, ref. 18059). Most of the spines are fi xed in the is an older name than C. antillarum Jordan and Rutter (1897, normal ʻburrfi shʼ manner, but those in the pectoral axil, which ref. 10644), and the Bloch and Schneider fi gure clearly applies are by far the longest on the body, are erectile. Some have to the same species, the name has been little used since its regarded Tragulichthys Whitley (1931, ref. 4673) as a synonym description other than as a junior synonym of either spinosus or or subgenus of Diodon (Fraser-Brunner, 1943, ref. 1495; Tyler, schoepfi i. Other than Paepkeʼs (1999, ref. 24282) catalogue of Bloch types (unfortunately, the type of geometricus is lost), the 1980, ref. 4477). But, until a full analysis of the phylogeny of the most recent correct use of geometricus was Günther (1870, family is forthcoming, it seems best to maintain current usage ref.1995), and this for only one of his ʻvarietiesʼ (i.e., beta). In and to recognize Tragulichthys at the generic level because the contrast, C. antillarum has been widely, almost universally, only species has a number of morphological differences from the used for this species (see Eschmeyer, 2005, for 13 publications fi ve species normally included in Diodon. between 1983 and 2003; in addition, Bailey et al., 1960, ref. 27285; 1970, ref. 27286; Böhlke and Chaplin, 1968, ref. 23150; Tragulichthys jaculiferus (Cuvier, 1818) Randall, 1968; 1996; Tyler, 1977; 1980, ref.4477; Robins et al., Diodon jaculiferus Cuvier, 1818: 130, pl. 7 (ʻIndian Ocean via 1980, ref. 7111; 1991, ref. 14237; Lieske and Myers, 1994; Peronʼ)* Cervigón, 1996, ref. 24489; Smith, 1997; Lyczkowski-Schultz, Chilomycterus grandoculis Ogilby, 1910: 19 (Moreton Bay, et al., 2005). Because the senior synonym (geometricus) has Queensland)* not been used as a valid name after 1899, and because the junior *extant type name has been used in at least 25 works published by at least Based on examination of 50 lots, including extant types, from ten authors over the last 50 years, this meets the criteria of throughout the range. Aside from confusion regarding the Articles 23.9.1 and 2 of ICZN (1999, ref. 26875), and prevailing designation of a type species for Allomycterus (see below), usage (of antillarum) must be maintained. In 2003, I suggested there have been few nomenclatural issues regarding this tropical that current usage of C. antillarum be maintained in the Australian species. References to this species from New interests of stability (Leis, 2003, ref. 27121), and here provide Zealand are of Allomycterus pilatus (see below): T. jaculiferus evidence that ICZN criteria require this to be met. does not occur in New Zealand. Chilomycterus orbitosus Poey (1875: 69, ref. 18564) is Cuvierʼs (1818, ref. 18059) description, fi gure, and the clearly a species of the “Atlantic Chilomycterus” group, but extant type leave no doubt about the identity of this distinctive there is no known extant type. Poeyʼs brief description on species. The description of Chilomycterus grandoculis Ogilby Cuban specimens seems to be based on a composite of C. (1910, ref. 3288) details the diagnostic spination, and the extant antillarum and C. schoepfi i from Cuba, but perhaps best fi ts type makes it clear that it is conspecifi c with T. jaculiferus. the former. The name orbitosus has not been used since its description, as far as I can ascertain. So, even if it could be Distribution. Northern Australia. established that orbitosus Poey, 1875 and antillarum Jordan From Derby, Western Australia to Darwin (including and Rutter, 1897 are conspecifi c, I would recommend against Rowley Shoals) to Torres Strait and south to Moreton Bay, the use of the older orbitosus in the interests of stability. Qld.

Distribution. W Atlantic. Dicotylichthys Kaup, 1855 Florida, Bahamas and Cuba to Venezuela, Barbados and Brazil Dicotlylichthys Kaup, 1855 (type species Dicotylichthys Some Brazilian specimens are intermediate in colour punctulatus Kaup) between the crisp, dark hexagonal pattern typical of C. Atopomycterus Bleeker (ex Verreaux), 1865 [type species Atopomycterus diversispinus Bleeker (ex Verreaux)] antillarum and the dark background with vague lighter spots typical of C. spinosus spinosus. The signifi cance of this is Diagnosis. Spines on head and belly erectile, those on back and unknown, and further investigation is required. sides fi xed; fi xed spines with 3 bases, erectile spines with 2 bases; spines long to medium; 9 C rays; 21 vertebrae; no tentacles; Tragulichthys Whitley, 1931 nostril in adult bifi d; no spines wholly on dorsal surface of caudal peduncle, but large spines extend over the peduncle nearly to the Tragulichthys Whitley, 1931 (type species D. jaculiferus Cuvier) caudal-fi n base; no fi ns spotted; no large blotches on dorsal Diagnosis. All spines fi xed, except those in pectoral axil which surface, but lateral bars present. Some additional osteological are by far the longest on the body; all spines except those in the characters are given by Tyler (1980, ref. 4477). P axil with 3 bases; spines long to medium; 9 C rays; 19 The type species of Dicotylichthys is D. punctulatus Kaup vertebrae; no tentacles in adults; nostril in adult a short tube (1855, ref. 2571). The sole species in this genus has erectile with 2 openings, but may become bifurcate in larger individuals; spines on the head and belly, but fi xed ones on the back and no spines wholly on dorsal surface of caudal peduncle, but sides. In contrast to the arrangement adopted here, some authors large spines extend over the peduncle nearly to the caudal-fi n follow Fraser-Brunner (1943, ref. 1495) and include all base; no fi ns spotted; no large blotches on dorsal surface. Some diodontids that develop bifi d nasal organs in Dicotylichthys. 86 Jeffrey M Leis

This would place in the same genus such disparate species as Diagnosis. All spines fi xed except 1 or 2 in P axil; spines with 3 pilatus with all fi xed spines, punctulatus with a mixture of bases except erectile ones; spines long or short; 9 C rays; no erectile and fi xed spines, and nicthemerus with all erectile tentacles in adults; nostril in adult bifi d; no spines wholly on spines. However, Dicotylichthys is very similar to the monotypic dorsal surface of caudal peduncle, but large spines that extend Lophodiodon (see below). Bleekerʼs (1865, ref. 416) description over the peduncle nearly to the caudal-fi n base; no fi ns spotted; no of Atopomycterus (based on an unpublished manuscript by large blotches on dorsal surface, but lateral bars may be present. Verreaux held in MNHN) is brief, but fortunately the types of The type species of this genus is A. jaculiferus (non- A. diversispinus Bleeker are extant (see below), thus clearly Cuvier) McCulloch. McCulloch (1921, ref. 2945), apparently showing that Atopomycterus is a synonym of Dicotylichthys. following Güntherʼs (1870, ref. 1995) concept of jaculiferus, provided an excellent illustration of the species he was Dicotylichthys punctulatus Kaup, 1855 proposing as the type of his new genus. However, it was not D. jaculiferus of Cuvier (1818, ref. 18059). Whitley (1931, ref. Dicotylichthys punctulatus Kaup, 1855: 230 (Cape of Good Hope, 4673) realized this, and also realized that McCullochʼs fi sh South Africa and Mauritius, but these localities are apparently was undescribed. Whitley therefore described as new incorrect, see below)* Allomycterus pilatus, and designated as his holotype the Atopomycterus diversispinis Bleeker (ex Verreaux), 1865: 49 (Australia)* specimen illustrated and described by McCulloch (1921). Not Dicotylichthys myersi Ogilby, 1910: 18 (Moreton Bay, Queensland, surprisingly, this has caused some confusion. Australia)* * extant type Allomycterus pilatus Whitley, 1931 Based on over 50 lots from throughout the range, including all Allomycterus jaculiferus (non-Cuvier) McCulloch, 1921: 141, pl. extant types. Kaupʼs (1855, ref. 2571) description is not detailed, 23 (fi g. 2) (New South Wales, Australia) but the extant specimens upon which he based his description Allomycterus pilatus Whitley, 1931: 125 (NSW, Australia)* Allomycterus whitleyi Phillipps, 1932: 13, fi g. 5 (New Zealand)* are all of this distinctive species. *extant type Although Bleekerʼs 1865 description of Atopomycterus diversispinis is brief, and not detailed, the syntypes are extant Based on 38 lots from throughout the range, including the and readily identifi ed as D. punctulatus Kaup (1855, ref. 2571; extant types. Confusion over the specifi c name of this species see Leis and Bauchot, 1984, ref. 12539). is dealt with under the genus. There seem to be two forms of Dicotylichthys myersi Ogilby (1910 ref. 3288) was said by this species, one with long, blade-like spines (A. whitleyi Ogilby to differ from D. punctulatus by the relative size of the form), and another with short, compressed spines. Both forms abdominal spines, but the syntypes are well within the range occur off the Australian mainland, but I have seen only the of relative spine size of D. punctulatus. long-spine form from New Zealand, and the specimens with the longest spines seem to be from New Zealand. These Distribution. South-eastern Australia. differences are not obviously connected with sexual From Moreton Bay, Qld to Bass Strait. dimorphism. Therefore, there may be two species of Kaup (1855, ref. 2571) reported that his type specimens came Allomycterus, and a genetic study would be useful in clarifying from the Cape of Good Hope and Mauritius. Subsequently, the situation. In addition, Kuiter (1993, ref. 23929) illustrates Günther (1870, ref. 1995) reported that the only specimen in two colour morphs among south-eastern Australian specimens BMNH identifi ed as being from Mauritius was of questionable of A. pilatus, referring to deep-water and shallow-water forms. locality, and that the sole specimen from the Cape of Good Hope The basis for the colour differences is unclear and should be (which he identifi ed as the ʻtypeʼ of D. punctulatus) was investigated. References to Allomycterus jaculiferus from “presented by Sir A. Smith”. Smith was a medical doctor resident New Zealand are based on A. pilatus (see discussion under in Cape Town who procured many specimens from passing Allomycterus). ships, and then provided them to the British Museum, where they Whitleyʼs (1931, ref. 4673) description and McCullochʼs were generally assumed to have originated in Cape Town (Bass (1921, ref. 2945) illustration are clear, and could apply to no et al., 1975, ref. 7409). Thus, there is good reason to question the other species. In addition, the holotype is extant. locality data of Smith specimens (Bass et al., 1975, ref. 7409) if Phillippsʼ (1932, ref. 16393) A. whitleyi constitutes the long- other evidence is inconsistent with them. Other than this BMNH spined form from New Zealand, and although both holotype and specimen, I have been unable to fi nd any institution (including paratype are stuffed and distorted, they appear to differ from A. RUSI) that has specimens of D. punctulatus from anywhere but pilatus only in the length and shape of the spines. Phillippsʼ Australia, where it is abundant within its range. Thus, I conclude description contains two spellings of the specifi c name: two as that Sir A. Smith procured his specimen from a passing ship, not whitleyi and one as whiteleyi. Given the correct spelling of from the Cape of Good Hope, and that this species is endemic to Gilbert Whitleyʼs name (to whom the patronym refers), “whiteley” south-eastern Australia. is clearly a typographical error even though it appears before the two uses of whitleyi within Phillipsʼ article. Allomycterus McCulloch, 1921 Distribution. Southern Australia and New Zealand. Allomycterus McCulloch, 1921 (type species Allomycterus Rottnest I., WA, to Botany Bay, NSW, including Tas.; jaculiferus [non-Cuvier] McCulloch = Allomycterus pilatus Whitley) Tasman Sea seamounts and ridges; and New Zealand. Species of porcupinefish 87

Lophodiodon Fraser-Brunner, 1943 Diodon eydouxii Brisout de Barneville, 1846 Lophodiodon Fraser-Brunner (type species Diodon calori Diodon bertolettii de Lema, de Lucena, Saenger and de Oliveira, Bianconi) 1979: 35–38, fi gs 18–19 (Brazil)* *extant type Diagnosis. Spines on head and belly erectile, those on back and sides fi xed; fi xed spines with 3 bases, erectile spines with 2 Diodon bertolettii can readily be identifi ed as a synonym of D. bases; spines short to medium; anteriorly-pointing spines on eydouxii Brisout de Barneville based on its semi-lunate fi ns, snout; 9 C rays; a small supraorbital tentacle in adults; nostril blue colour, fi n-ray counts, and from the photographs provided in adult a short tube with 2 openings; no spines wholly on by de Lema et al. (1979, ref. 8836). dorsal surface of caudal peduncle, but large spines extend over Leis (1978, ref. 5529) examined specimens of D. eydouxii the peduncle nearly to the caudal-fi n base; no fi ns spotted; no Brisout de Barneville (1846, ref. 296) from 19 scattered large blotches on dorsal surface, but bars present laterally. localities in all warm oceans. I have now seen an additional 31 The type species of this genus is D. calori Bianconi (1854, lots. These plus four acceptable literature records extend the ref. 17949). The sole species in this genus has most spines on known distribution of this species. head and belly erectile, and those on back and sides fi xed. This genus is similar in many ways to Dicotylichthys, differing Distribution. Pelagic, Atlantic, Indian and Pacifi c Oceans. primarily in that the nasal organ in Dicotylichthys is bifi d, W Indian – Cape of Good Hope to Zanzibar whereas in Lophodiodon, it is a hollow tube with two distinct E Indian – only record is Andaman Sea nostrils. There is reason to expect that two genera may W Pacifi c – Indonesia to Okinawa eventually be considered to be synonymous. Central Pacifi c – from near Samoa to Hawaii E Pacifi c – equator to 20oN plus a California record (Lea, Lophodiodon calori (Bianconi, 1854) 1998: from Los Angeles Harbour, but misidentifi ed as C. reticulatus [Linnaeus]). Diodon calori Bianconi, 1854: 69 (Mozambique) W Atlantic – 28oS to 37oN Lophodiodon nigropunctatus Smith, 1957: 222, fi g. 4 (Port E Atlantic – 30oW is eastern-most specimen examined, but Alfred, South Africa)* *extant type there are apparently valid literature records from the Azores (Azevedo, 2004) and from Spain (Crespo et al., 1987). Based on 13 specimens from most of the range, including the extant types. This species is widely distributed, but uncommon Diodon holocanthus Linnaeus, 1758 in collections. Although the name Diodon calori dates from Bianconi, Diodon paraholocanthus, Kotthaus, 1979: 39, fi g. 492 (Bab-el- 1854 (ref. 17949), the illustration of Diodon calori in Bianconi Mandeb, southern Red Sea)* *extant type (1855, ref. 295) is diagnostic for this species, with its large number of short spines, four lateral bars and no dorsal Kotthaus (1979, ref. 8818) confused Diodon liturosus Shaw blotches. with Diodon holocanthus Linnaeus, as is obvious from his Lophodiodon nigropunctatus Smith (1957, ref. 12171) was description and photograph (his fi g. 491) of what he called D. based on juveniles (30–60 mm SL), and the apparent difference holocanthus. Then, having encountered the true D. holocanthus in colour pattern with L. calori can be attributed to this. The in the north-western Indian Ocean, he described it as a new spination of Smithʼs specimens is diagnostic. species, D. paraholocanthus. The description and photo (his Distribution. Indo-Pacifi c. fi g. 492) of the holotype are entirely consistent with D. The 13 specimens I have examined are all from east Africa holocanthus Linnaeus. and Seychelles, but the species is reliably reported from Oman, Leis (1978) examined 141 specimens of D. holocanthus Bali, Timor, the Australian Northwest Shelf, the South China Linnaeus (1758, ref. 2787) from all warm oceans. I have now Sea and New Caledonia, and somewhat less reliably as the seen more than 100 additional lots that extend the known similar D. punctulatus from New Guinea by Tortonese (1964, distribution of the species (see below). It is noteworthy that ref. 9080) and Munro (1967, ref. 6844). there are still no records of D. holocanthus from the Pacifi c Plate other than those reported by Leis (1978): Hawaii, Easter Diodon Linnaeus, 1758 and Pitcarin Is. Reference in Robertson et al. (2004) and Mundy (2005, ref. 28379) to D. holocanthus occurring in the This genus was revised by Leis (1978, ref. 5529) with additional Line Is. is incorrect (B.C. Mundy, personal communication). information on nomenclature and types in Leis and Bauchot (1984, ref. 12539), and information contained there is not Distribution. Circumtropical in Atlantic, Indian and Pacifi c repeated. Only information on Diodon species described since Oceans (except only peripherally on Pacifi c Plate). 1978 and on noteworthy new distributional information is W Indian – from Cape of Good Hope, South Africa to included here. Note that fi gs 9 and 17 of Leis (1978) were Oman and Red Sea, Sri Lanka, Mascarenes, and Seychelles. switched (see 1979 errata facing p. 956, US Fishery Bulletin E Indian – Andaman Sea to Australia 76[4] ): fi g. 9 labelled Diodon hystrix is actually D. holocanthus W Pacifi c – west of Pacifi c Plate: Japan to New Caledonia and fi g. 17 labelled Diodon holocanthus is actually D. hystrix. and Elizabeth and Middleton Reefs, Tasman Sea. 88 Jeffrey M Leis

In Australia, south to Ulladulla, NSW (36oS) off east coast, Key to genera and species of the family Diodontidae and to Freemantle, WA (32oS) off west coast. NB: in juveniles relative spine length and body colour generally Central Pacifi c – Hawaii, Easter and Pitcarin Is. only. differ from those of adults E Pacifi c – southern California to Colombia W Atlantic – Hudson Canyon (off New Jersey) to 1. All body spines erectile and 2-rooted (except a few around Argentina. gill opening or dorsal-fi n base) Diodon 10 E Atlantic – Liberia and Nigeria to northern Angola . — All or most body spines of back and sides fi xed in an erect position and 3-rooted 2 Diodon hystrix Linnaeus, 1758 Non-Diodon 2. Indian and Pacifi c in distribution 3 Leis (1978) examined 43 specimens of Diodon hystrix Linnaeus — Atlantic in distribution (1758, ref. 2787) from all warm oceans. I have now seen an (NB: 1 Indo-Pacifi c species, Cyclichthys spilostylus, has additional 80 lots that extend the documented distribution of penetrated the eastern Mediterranean Sea through the the species (see below). Suez Canal 14 Distribution. Circumtropical in Atlantic, Indian and Pacifi c Indo-Pacifi c non-Diodon Oceans 3. Spines on top of head and on belly erectile 4 W Indian Ocean – throughout the area from South Africa — Spines on top of head and on belly fi xed in an erect (Tsitsikamma Coastal National Park) to the Red Sea, Sri position 5 Lanka, and all major island groups. 4. 2 to 4 spines in the 1st row on the snout point toward the Australia – south to Elizabeth and Middleton Reefs, Lord mouth when not erect; no small, black spots scattered more Howe I., and northern NSW (29oS) off the east coast, and or less uniformly over head and trunk Rowley Shoals on west coast. Lophodiodon calori (Indo-west Pacifi c) W Pacifi c – New Caledonia and Kermadecs to Rotuma, — All erectile spines point toward tail when depressed; small, black spots scattered more or less uniformly over Pitcarin I., Hawaii and southern Japan. head and trunk E Pacifi c – Mexico to Chile o o Dicotylichthys punctulatus (south-east Australia) W Atlantic – 36 N to ca 20 S 5. A small spine or 2 wholly on the dorsal surface of the Central Atlantic – Ascension and St Helena caudal peduncle; normally 10 caudal rays; nasal organ of E Atlantic – only 1 confi rmed record at Fernando Po adults an open ridged cup; adults with fi ns spotted Chilomycterus reticulatus (circumtropical) Diodon liturosus Shaw, 1804 — No spines wholly on the caudal peduncle; normally 9 caudal rays; nasal organ of adults a short tube with either Leis, 1978 examined 30 specimens of Diodon liturosus Shaw 2 openings or split at the end (not an open cup); no spots (1804, ref. 4015) primarily from the western Pacifi c. I have on fi ns of adults 6 now seen an additional 45 lots that extend the documented 6. A set of 4 long fi xed spines with their bases near the dorsal distribution of the species (see below). and anal-fi n bases – their pointed ends extend over the Distribution. Indo-west Pacifi c caudal peduncle; a few spines in P axil erectile 7 W Indian Ocean – South Africa (Algoa Bay) to Oman and — No especially long spines around dorsal and anal fi n southern Red Sea, Mascarenes, Seychelles, Laccadives and bases; spines in P axil fi xed 8 7. Very long spines (longer than rays of pectoral fi n) in Maldives. pectoral-fi n axil; 3–4 black spots (< eye) on sides of head E Indian Ocean – Phuket, Thailand to Ningaloo Reef, and trunk, none on back WA Tragulichthys jaculiferus (tropical Australia) W Pacifi c – from Maizuru, Japan to northern NSW, — Spines of pectoral-fi n axil not particularly elongate; some Australia to New Caledonia to Society and Marshall Is. eye-size dark spots on back generally associated with spine bases Allomycterus pilatus Diodon nicthemerus Cuvier, 1818 (temperate Australia, Tasman Sea and New Zealand) Leis (1978) examined nine specimens of Diodon nicthemerus 8. Few black spots on body, those present at base of spines dorsally and dorso-laterally; D, A and C fi ns with dusky Cuvier (1818, ref. 18059), all from southern Australia. Museums distal margin; only 2 spines over eye; 2 spines between in Australia contain large numbers of this species, and its nostrils, 1 immediately adjacent to each nostril distribution is confi rmed as being confi ned to the waters of Cyclichthys hardenbergi southern Australia. I have seen specimens from an area ranging (tropical Australia, southern New Guinea) from Houtman Abrolhos Is., WA (28oS), to Nadgee, NSW — Black spots in clusters dorsally and laterally, or associated (37oS), although Kuiter (1993) reports D. nicthemerus as far with spine bases laterally and ventrally; D, A and C fi ns north as Seal Rocks (32oS). This is the most restricted either clean or with faint, parallel bands; 3 spines over distribution of any species of Diodon. eye; only 1 spine between nostrils, located medially 9 Species of porcupinefish 89

9. Spines few, 4 dorsally between pectoral-fi n bases, 8 or 9 Atlantic non-Diodon anterior to dorsal-fi n base; a short, moveable spine near 14. 1 or 2 small spines wholly on the dorsal surface of the corner of mouth; all spines on top of head with 3 bases; caudal peduncle; normally 10 caudal-fi n rays; nasal organ black spots in clusters dorsally and dorsolaterally of adults, an open, ridged cup; adults with fi ns spotted; on Cyclichthys orbicularis (Indo-west Pacifi c) top of head some spines with 4 roots — Spines more numerous, 5 or 6 dorsally between pectoral- Chilomycterus reticulatus (circumtropical) fi n bases, 11 or 12 anterior to dorsal-fi n base; no moveable — No spines wholly on the caudal peduncle; normally 9 spines; some spines on top of head with 4 bases; black caudal-fi n rays; nasal organ of adults, a short hollow tube spots at base of spines laterally and ventrally with 2 openings; fi ns of adults usually without spots; all Cyclichthys spilostylus (Indo-Pacifi c) spines with 3 roots “Atlantic Chilomycterus” 15 15. A large (ca. = eye diameter) tentacle above eye; colour Diodon pattern dominated by large dorsal blotches and with small 10. None of spines wholly on caudal peduncle; body with spots scattered on back and sides, on fi ns only basally, several large, dark dorsal or lateral blotches; no small, except on most or all of caudal fi n from 10–15 cm standard dark spots on fi ns 11 length, and on other fi ns from 20 cm — One or more small spines wholly on the dorsal surface of Chilomycterus antennatus (central-west Atlantic) caudal peduncle; body without large dorsal blotches; all fi ns — Tentacles above eyes absent or small; no small spots on (anal sometimes excepted) heavily spotted 13 fi ns or on back and sides; dorsal and lateral dark blotches 11. Temperate Australian waters only; no small, fi xed, tri- present 16 base spine immediately above gill opening; no small, fl at 16. Network of hexagonal to circular black lines on back and spines on the anterior border of the depression surrounding sides in adults the gill opening; 11 or fewer spines from lower jaw to Chilomycterus antillarum (central-west Atlantic) anus; adult colour pattern dominated by 4, large, lateral — Black lines on back and sides absent in adults, or if present, bars, dorsum uniformly dark wavy or approximately parallel – not intersecting to form Diodon nicthemerus (southern Australia). rings or polygons 17 — Tropical waters, with strays into warm temperate water; 1 17. Extensive series of dark-brown to black parallel lines or 2 small, fi xed tri-base spines above gill opening; 3 or 4 densely covering back and sides in adults small, fl at spines forming the anterior border of depression Chilomycterus schoepfi i (western North Atlantic) surrounding the gill opening; 12 or more spines from — No black lines on back; dorsal background dark with diffuse lower jaw and anus; adult colour pattern dominated by lighter spots Chilomycterus spinosus 18 several large, dorsal blotches 12 18. No black lines on sides of head or trunk; South American 12. Frontal spines obviously much shorter than spines in distribution Chilomycterus spinosus spinosus immediately behind pectoral fi n; small downward- (east coast South America) pointing spine below anterior margin of eye; 17–22 spines — Irregular, approximately parallel black lines on sides of from lower jaw to anus; large dorsal blotches with distinct head and trunk; eastern Atlantic in distribution pale border; blotch below eye not continuing over top of Chilomycterus spinosus mauretanicus head Diodon liturosus (Indo-Pacifi c) (west coast Africa) — Frontal spines slightly shorter to much longer than spines immediately behind the pectoral-fi n base; small Acknowledgements downward-pointing spine below anterior margin of eye absent (Indo-Pacifi c) or present (most Atlantic specimens); I thank the curators and staff of the many Museums who have 12–15 spines from lower jaw to anus; dorsal blotches allowed me access to their collections of diodontids over many without distinct pale border; blotch below eye continues years and to many colleagues who provided information on over interorbital in Indo-Pacifi c specimens diodontids. W.N. Eschmeyer, CAS, and J.C. Tyler, USNM, Diodon holocanthus (circumtropical) offered constructive comments on a draft of the manuscript. 13. Pectoral- fi n rays 19–22; anal-fi n rays 16–18; dorsal and S. Bullock provided editorial assistance. This research was anal fi ns somewhat pointed to semilunate in adults; supported, over many years, by the Australian Museum, Sydney. relatively streamlined, head width of adults 3.3–4.0 in standard length; 10–14 spines from lower jaw to anus; a References wholly pelagic species coloured dark-blue dorsally Publications whose citations contain a ref. number can be found Diodon eydouxii (circumtropical) in Eschmeyer, 2005, and are not repeated here. — Pectoral- fi n rays 22–25 (rarely 21); anal-fi n rays 14–16; dorsal and anal fi ns rounded in adults; relatively robust, Artedi, P. 1738. Ichthyologia (Lugduni Batavorum). Pars III Genera Piscium. 1962 reprint, Hostorae Naturalis Classica 15, J. Cramer: head width of adults 2.4–3.3 in standard length; 14–19 Weinheim, pp 203–294. spines from lower jaw to anus; juveniles (up to 20 cm) Azevedo, J.M.N., Raposeiro, P.M. and Rodrigues, L. 2004. First pelagic and coloured blue dorsally, adults demersal and records of Fistularia petimba and Diodon eydouxii for the Azores, coloured tan to brown with notes on the occurrence of three additional species. Journal Diodon hystrix (circumtropical) of Fish Biology 65: 1180–1184. 90 Jeffrey M Leis

Crespo J., Rey, J.C. and Garcia, A. 1987. Primera cita de Acanthurus (eds) FAO species identifi cation sheets for fi sheries purposes: the monroviae Steindachner, 1876 y de Diodon eydouxii Brissout de Western Indian Ocean. FAO: Rome. 8 pp (unpaginated). Barneville, 1846 para la ictiofauna europa. Miscellània Zoològica, Leis, J.M. (in press) Family Diodontidae in: Carpenter, K.E. (ed) The 11: 271–275. living marine resources of the Eastern Central Atlantic, FAO, Eigenmann, C.H. 1885. A review of the genera and species of Rome Diodontidae found in American seas. Annals of the New York Lieske, E. and Myers, R. 1994. Coral reef fi shes. Indo-Pacifi c and Academy of Science. 3: 297–311. Caribbean including the Red Sea. Harper Collins, 400 pp Eigenmann, C.H. 1892. On the occurrence of the spiny boxfi sh (genus Lozano Rey, D.L. 1952. Peces Fisoclistos. Memorias de la Real Chilomycterus) on the coast of California. Proceedings of the Academia de Ciencias Exactas, Físicas y Naturales de Madrid. United States National Museum 15: 485, plate LXXXI. 14: 1–378. Eschmeyer, W.N. 2005. The catalogue of fi shes on-line (17 October Lyzkowski-Schultz, J., Zapfe, G.A. and Bond, P.J. 2005. Diodontidae: 2005 version). http://www.calacademy.org/research/ichthyology/ Porcupinefi shes, pp 2449–2455 in: Richards, W.J. (ed) Early catalog/fi shcatsearch.html stages of Atlantic Fishes: an identifi cation guide for the western Golani, D. 1993. Trophic adapatation of Red Sea fi shes to the eastern central Atlantic. Taylor & Francis: Boca Raton. Mediterranean environment – review and new data. Israel Journal Randall, J.E. 1968. Caribbean reef fi shes. TFH Publishing, Jersey of Zoology 39: 391–402. City. Heck, K.L. and Weinstein, M.P. 1978. Mimetic relationships between Randall, J.E. 1996. Caribbean reef fi shes, 3rd edn. TFH Publishing, tropical burrfi shes and opisthobranchs. Biotropica 10(1): 78–79. Hong Kong. Humann, P. 1997. Reef fi sh identifi cation: Galápagos. New World Robertson, D.R., Grove, J.S. and McCosker, J.E. 2004. Tropical Publications: Jacksonville, Florida transpacifi c shore fi shes. Pacifi c Science 58 (4): 507–565. Kailola, P.J. 1975. A catalogue of the fi sh reference collection at the Séret, B. and Opic, P. 1981. Poissons de mer de lʼOuest Africain tropical. Kanudi fi sheries research laboratory, Port Moresby. Papua New Initiations-Documentations Techniques 49, ORSTOM: Paris. Guinea Department of Agriculture and Fisheries Research Smith, C.L. 1997. National Audubon Society fi eld guide to tropical Bulletin, 16: 1–277. marine fi shes. Chanticleer Press: New York Lea, R.L. 1998. The spotfi n burrfi sh – an infrequent visitor. Tidelines Tyler, J. C. 1977. Diodontidae in: Fisher, W. (ed) FAO species 18(1): 8. identifi cation sheets for fi sheries purposes: the Western Central Leis, J.M. 1981. Family Diodontidae in: Fisher, W. (ed) FAO species Atlantic Ocean. FAO: Rome. 4 pp (unpaginated) identifi cation sheets for fi sheries purposes: the Eastern Atlantic Whitley, G.P. 1952. Porcupine fi shes. The Australian Museum Ocean. FAO: Rome. 4 pp (unpaginated). Magazine. 10 (11): 353–360. Leis, J.M. 1984. Family Diodontidae in: Fisher, W. and Bianchi, G. Willughby, F. 1686. Historia piscium libri quatuor. Oxonii